Thirty-day readmissions among patients with cardiogenic shock who underwent extracorporeal membrane oxygenation support in the United States: Insights from the nationwide readmissions database

Machine learning based predictive modeling of readmissions following extracorporeal membrane oxygenation hospitalizations

“…In our sensitivity analysis which solely utilized patient-level factors as readmission predictors, XGBoost continued to exceed LR across performance metrics. In line with research evaluating the impact of socioeconomic disadvantage on outcomes following ECMO for cardiogenic shock, we found lowest income and public insurance to confer greater readmission risk [ 5 ]. Furthermore, we identified patient frailty and comorbid liver disease as independent risk factors for rehospitalization.…”

“…Current models, however, remain limited due to the use of older, highly selective, or pediatric cohorts, thus reducing generalizability without accounting for local variations in clinical practice. Furthermore, published models generally lack external validation and demonstrate poor predictive performance, limiting their clinical utility [ 5 ]. Machine learning (ML) algorithms which utilize nonlinear data structures hold the promise of yielding prediction models with superior performance relative to traditional regression methods [ 6 ].…”

Contemporary approach to cardiogenic shock care: a state-of-the-art review

“…The CSWG demonstrated that in-hospital mortality was significantly higher in patients with AMI-CS (39.5%) as compared to HF-CS (25.3%; P < 0.0001) despite having similar hemodynamic profiles ( 43 ). In a Nationwide Readmission Database study, there was a 16% readmission rate among 4,229 survivors of CS post-ECMO who recovered and were discharged alive ( 160 ). These patients had an in-hospital mortality rate of 10% with the most common cause of re-admission being infection followed by acute decompensated heart failure ( 160 ).…”

“…In a Nationwide Readmission Database study, there was a 16% readmission rate among 4,229 survivors of CS post-ECMO who recovered and were discharged alive ( 160 ). These patients had an in-hospital mortality rate of 10% with the most common cause of re-admission being infection followed by acute decompensated heart failure ( 160 ). Additional CS studies analyzing patients with AMI-CS and non-AMI-CS (not restricted to ECMO use), respectively, reported a higher readmission rate of 20% and 23%, respectively ( 161 , 162 ).…”